81 World Health Organization who technical Report Series, No. 957, 2010
Download 176.21 Kb. View original pdf
|
| Evaluation of test results 18.1 Test results should be reviewed and, where appropriate, evaluated statistically after completion of all the tests to determine whether they are mutually consistent and if they meet the specifi cations used. The evaluation should take into consideration the results of all the tests (all test data. Whenever doubtful (atypical) results are obtained they should be investigated. The complete testing procedure needs to be checked according to the internal quality management system see also Part one, section 2). 18.2 When a doubtful result (suspected OOS result) has been identifi ed, a review of the different procedures applied during the testing process is to be undertaken by the supervisor with the analyst or technician before retesting is permitted. The following steps should be followed: (a) con rm with the analyst or technician that the appropriate procedures) was (were) applied and followed correctly; (b) examine the raw data to identify possible discrepancies; (c) check all calculations; (d) check that the equipment used was qualifi ed and calibrated, and that system suitability tests were performed and were acceptable; (e) ensure that the appropriate reagents, solvents and reference substances were used; (f) con rm that the correct glassware was used and (g) ensure that original sample preparations are not discarded until the investigation is complete The identifi cation of an error which caused an aberrant result will invalidate the result and a retest of the sample will be necessary. TRS957.indd 116 21.04.10 11:03 117 Doubtful results can be rejected only if they are clearly due to an identifi ed error. Sometimes the outcome of the investigation is inconclusive — no obvious cause can be identifi ed — in which case a con rmatory determination is to be performed by another analyst who should beat least as experienced and competent in the analytical procedure as the original analyst. A similar value would indicate an OOS result. However, further con rmation using another validated method, if available, maybe advised An SOP should be in place for the conduct of an investigation of an OOS test result. The SOP should give clear guidance on the number of retests allowed (based on sound statistical principles. All investigations and their conclusions should be recorded. In the event of an error, any corrective action taken and any preventive measure introduced should be recorded and implemented All individual results (all test data) with acceptance criteria should be reported All conclusions should be entered on the analytical worksheet (see Part three, section 15) by the analyst and signed by the supervisor. Note: Further guidance on evaluation and reporting of test results is available in the following Guideline elaborated by the US Food and Drug Administration (5); • Guideline elaborated by the European Network of Of cial Medicines Control Laboratories (OMCL) (28). Download 176.21 Kb. Share with your friends:
|